Saturday, January 11, 2014

Culture and A Patent's Multiple Inventors

PatentlyO yesterday put up a new chart and discussion about the steady increase in the number of multiple inventor inventions, and the correlating decrease in the number of single inventor inventions (see also PatentlyO's previous post on this topic: suggesting the increase has been happening at least since the 1970s). As a patent blog, of course, “invention” here stands directly for “patent” and only indirectly for innovation:

This image is copyrighted by Dennis Crouch, Patently-O and is available at http://www.patentlyo.com.

In his post, Dennis Crouch laments the slow but steady loss of the original, solitary inventor that this portends. He starts the post with a reference to the way that (American?) culture emphasizes extroversion over introversion and then ties this to the increase in teamwork centered patents. While I believe the comment on culture increasingly valuing extroversion may have some merit to it, at least as limited to the United States, the link between this and multiple inventor patenting is not quite so easy to make.

What about the portion of US patents that are not invented in the United States? The proportion of this type of foreign application at the USPTO has been increasing over the last few decades, now taking up roughly 50% of total applications. They are therefore not just more and more important in terms of the overall make-up of granted US patents, but they are also more and more numerically significant in terms of calculating statistical trends. Do these types of applications also show an increase in multiple inventor inventions? My guess is that they do. While some may suggest that this is a result of some sort of overarching culture of modernity, this is unlikely given both the number of different cultural groups this would entail moving in the same direction and the ongoing dramatic differences between these groups in patenting styles. To boot, the vast majority of anthropological research into “modernity” has emphasized an increasing preoccupation with “the individual,” certainly not a trend that is easily fit with a parallel increase in multiple inventor patents. It could be that our preoccupation with the individual has caused us to ask this kind of question and therefore to “see” (and find significant) such a loss of the individual inventor. I don't, however, buy this response either. I believe the trend really is there. Since the United States has long celebrated a romantic view of a lone, genius (though perhaps a bit crazy) inventor it makes little sense to argue that we would only be recognizing this trend today.

So going back, then, to the argument that there may be a wide-ranging shift in cultures towards extroversion and therefore teamwork in patents, in my own fieldwork with Taiwanese patent and process engineers, I was told on several occasions of significant cultural differences in patenting practices (see also Libecap, 1989, "Contracting for Property Rights" on the lack of predicted movement toward some sort of global economic optimum in patent laws themselves). For instance, I was told that the Japanese style was to patent every little advance made in its own separate application. Americans, on the other hand, tended to wait until their smaller advances had added up to something slightly larger, a clear step forward rather than incremental improvement, before they would submit a new patent application. This particular difference in patenting behavior means that it makes little sense to compare the raw number of Japanese patents to the raw number of American patents. There are, of course, many further differences; this is only one example. Given such differences in styles of patenting, it would be quite surprising to see an overall move, based on a culture of modernity, toward multiple inventor patents.

This example, however, is also a good example of an odd trend among non-anthropologists. People (and I mean this in terms of Americans as well as Chinese and Taiwanese, at least) seem to have become more and more comfortable resorting to “culture” as an explanation of behavior. It is a quick, readily available, explanation that from an anthropologist's perspective often fails to really explain anything. Perhaps ironically, as we anthropologists are supposed to be culture's academic champions, what I am suggesting is that blaming culture has become too easy. Often, upon further analysis, cultural explanation turns out to have been masking other reasons for behavior. It would, for instance, make more sense to suggest that an overall change in patenting behavior at the USPTO is linked to policy at the USPTO rather than having its origins in simultaneous global cultural change. If this type of explanation based on American law, American legal precedent, or other patenting trends at the USPTO were insufficient, then it makes sense to look beyond this to overall trends. Going back to the example of Japanese versus American patenting differences, for instance, though these were relayed to me as cultural differences, they may not really be cultural at all. When the Company planned on patenting their Taiwanese inventions in Japan, for instance, they too would work on writing the patent in a way that fit with the larger Japanese patenting trend. For the US, several of these patents may be brought over together (see my earlier blog discussion of Nakamura's Japanese LED patent) into a single application (and then multiple divisionals as required by the patent examiner). Patenting style seems to me to be much more closely related to the rules and structures of that nation's patent office and courts of law1 rather than something more like “culture” per se.

So then where does this leave us? As a qualitative rather than quantitative researcher of patents, this is a question that I cannot answer definitively. However, what I can do is offer potential alternative solutions that quantitative researchers may not have thought of (or may not have thought significant enough). At the Company, one of the things that I noticed was that there were almost never patents submitted to any nation's PTO with only a single inventor. This is, in part, a matter of the nature of LED inventions: they require work on very expensive machines and they tend to build directly on work done prior to them. In my dissertation I discuss the creation of new knowledge in the Company's R&D department as a matter of working within a machine-material system. Any change you make within one layer of an LED chip (think of a layer “cake”) will, at the very least also have a direct impact on the physical connection, the electrical flow, and the flow of produced light between that layer and the layer it is grown upon. It will also affect the layer that will be grown on it. It may further affect the kinds of processes that can or now cannot be done many steps later in the production process (the new layer, may for instance, be more sensitive to temperature therefore forcing the engineers to turn up the pressure and down the temperature for a later, seemingly unrelated stage). All of this is to say that any particular new idea that gets implemented within an LED structure will involve input from a large number of people in charge of other portions of the production of that product. At least some of these will inevitably involve changes to the “invention” itself that will find their way into the patent's specification or claims. The complicated manufacturing techniques involved in LEDs (as well as in other semiconductor products) means that more people need to be consulted. The expensive nature of their production means that innovation in this field is extremely unlikely to come outside of a company or large research group-style academic lab. Nakamura's own early work on LED blue (and the surprise it gave to the industry) is the exception that proves the rule. Taking this a step beyond my own research, has there been an overall increase in the proportion of patents that come from this type of industry within the overall patent archive? Given the importance of semiconductor technologies and pharmaceuticals, at least, I would guess that their rise as complicated, high capital industries would account for some of the shift toward multiple patenting. I also would guess that, at least beyond the early Nakamura style work, the proportion of these industries' patents that have had multiple patents will have stayed relatively high.

A second place to look for why it is that multiple inventor patents have been increasing is the importance of the stage that an industry is in. Early on in a product's lifecycle improvements may be rather large and, perhaps, relatively easy to come by given the application of significant resources. In LEDs, early improvements would raise the efficacy of the products by huge percentages. By the time of Taiwan's early LED industry (which began in the 1970s, a decade or so after its beginning in the US and Russia), improvements were rather more modest in scale, perhaps on the order of 20-50 percent improvement at a time. By the time I arrived in the Company, they were starting to value improvements that gave maybe only a 3-4 percent improvement in efficacy; these were improvements that only a few years earlier would not have been thought of in terms of patents (or perhaps not even noticed) until they added up to around 10 percent.

It is not that innovation has slowed in the industry, however. The problem is exactly that there has been so much product innovation that meaningful competitive differences between products may now be quite small in percentage terms. In the beginning, when there were tremendous percentage increases in brightness, this was due to the fact that early LEDs were so dim. Now, there are considerably more patents in the LED industry each year than there were before. More and more companies seek to protect themselves from proliferating lawsuits with defensive portfolios while also searching for anything that might give them an advantage in terms of negotiating a cross-licensing deal with any of the LED industry's Big 5. As a result, companies increasingly need patents and their need for patents goes above, beyond and is separate from their need for product innovation.

However, in order to get a patent issued in the United States, you must show a significant improvement over the patents already in the USPTO archive. To do this consistently, and to avoid spending money on a patent application they might later have to abandon, the Company's patent engineers might combine several internal invention ideas into a single application. In terms of patent law, this means that the creation of the patent's “single invention” now happens within the IP department much more than it does in R&D. As a direct result of the merging of multiple related (at least in the terms of patents if not in the terms of the lab) ideas into a single invention, these patents will always have more than one inventor. I suggest that this experience with “merging” is not something unique to the Company, but rather something that makes sense due to the rules of the USPTO (and other national PTOs) and the particular competitive situation of the LED industry. These same structural conditions are not difficult to then find in other industries as well. Returning to the original question of the number of multiple inventor patents, remember that this is a measure, not of the number of multiple inventor ideas in the lab, but of the number of multiple inventor patents that issue. If early patents in an industry tend to have single inventors and later ones multiple inventors, then the fact that more mature industries have much more patenting activity overall means that the overall numbers will tend to be skewed toward multiple inventor patents as the proportion of relatively mature industries rises within the overall set of patenting industries. As you get more patents (which we have), in many cases you will also get even more multiple inventor patents.

Finally, the overall increase in multiple inventor patents might also be explained by a relatively steady increase in the emphasis the USPTO gives to inventorship. A couple of comments on the original post (including one by Patent_Guru) point out how important it is for companies (or other applicants) to get the exact inventors “right.” The IP engineers in the Company had a responsibility to the Company to ensure that anyone listed on the patent was an actual inventor and that every inventor, no matter their rank, was actually listed on the patent. Mislabeling (ie. misrepresenting) either way could result in a powerful patent being invalidated at a future infringement trial or a potentially dangerous witness's testimony. While there were therefore many different sized contributions to a patent, all of the names were partial inventors and no one could be added just because they built the original platform or were the inventors' boss. If there has been a growing emphasis on precision in inventorship, this might also partially explain the increase in multiple inventors.

The increase, I suggest, is therefore less attributable to an overall change in global (or even American culture) and more to, first, a change in the predominance of a different type of industries and, second, to a difference in patenting trends for early as opposed to more mature industries. A the third, overall shift in emphasis within US patent law and court decisions may also have played a role in companies listing anyone who might have contributed just to be safe (perhaps less dramatic than the shift in US patenting style towards listing more and more references). What I find particularly interesting, however, is not the increase in multiple inventors or decrease in single inventors, but the fact that double inventor inventions have stayed steady. Is this a matter of the graph being expressed in percentages rather than absolute numbers? Is “unchanging” or rather “changing equally up as down” (ie. What might have been listed with a single inventor in the past now has triple inventors or more, but a replacement number of what would have been listed as single inventor patents in the past are now listed as double)?

Monday, November 4, 2013

Chen Shuibian's Green Silicon Island Dream

The following is a graphic I put together to illustrate President Chen's "Green Silicon Island" campaign theme. Chen's embrace of environmentalism along with his actual election in 2000 marked a transition for the environmental movement in Taiwan. Prior to this, concern over the environment also meant opposition to the government. Following Chen's election and his appointment of environmental activists to the Taiwan Environmental Protection Agency and a few other environment-oriented agencies, environmentalists were presented with somewhat of a dilemma. Though they held key positions, they did not have the power to change the structural position of the DPP government nor to greatly influence DPP economic development policies. The DPP for its part, headed a government that never held a majority in the legislature and, slammed by an economic crisis and threats by Taiwanese companies to move to China, soon found itself beholden to the same industries as the KMT had been (but without the KMT's influence). Environmental concerns were, again, pushed to the back burner and activists found it difficult to mount resistance to a party that they had grown close to.

On building a Green Silicon Island, excerpt from Chen Shui Bian's Presidential Inauguration Speech, May 20, 2000:


"Today, facing the impact of the fast-changing information technologies and trade liberalization, Taiwan's industrial development must move toward a knowledge-based economy. High-tech industries need to be constantly innovative, while traditional industries need to undergo transformation and upgrading. […] Apart from consolidating our democratic achievements, promoting government reforms, and raising economic competitiveness, the new government's foremost objective should be to adhere to public opinion and implement reforms, so that the people on this land can live in more dignity, more self-confidence and better quality. […] The government will have to bring up solutions for all issues relating to the people's lives, such as social order, social welfare, environmental protection land planning, waste treatment, cleaning up rivers and community-building. It will also have to implement these solutions thoroughly. At present, we need to immediately improve social order and environmental protection, which are important indicators of the quality of life. Building a new social order, we will let the people live and work in peace and without fear. Finding a balance of ecological preservation and economic development, we will develop Taiwan into a sustainable Green silicon island."


The graphic uses, in part, creative commons licensed (or public domain) photographs from the Wikimedia Foundation (Formosan Map), Wade Brooks (flickr, LED bulb), yellowcloud (flickr, microchip), Photo Mojo Mike (flickr, solar panels), and james4765 (flickr, LED chips on board). Check out Ho Ming Sho's 2005 and 2010 works for more on the changes in Taiwan's environmental movement following Chen's election:

Ho 2005: Weakened State and Social Movement: The Paradox of Taiwanese Environmental Politics after the Power Transfer

Ho 2010: Environmental Movement in Democratizing Taiwan (1980–2004): A Political Opportunity Structure Perspective



Creative Commons License
This graphic (most recently) by Matthew West is licensed under a


Friday, September 27, 2013

NATSA 2012 Talk on Patents and the Taiwanese LED Indutsry

I just found a video of a talk I gave at the University of Indiana-Bloomington for the North American Taiwan Studies Association's 2012 Conference on Patents and the Taiwanese LED Indutsry. There are links to videos of all of the 2012 panels on NATSA's facebook page. I've edited this clip to include only my own talk and subsequent discussant remarks from Derek Sheridan (Brown University) and Joseph Wong (University of Toronto). There is a bit of background buzz, but all most all of talk is audible on the recording.

Monday, July 1, 2013

Property as Process: Nakamura's '839 Patent Family and What it Suggests about a Patent's Relationships to Time

A few months ago, I was doing some research into a particular LED patent issued to Shuji Nakamura (and assigned to Nichia Chemicals, his Japanese then-employer). This was one of the early blue LED patents (GaN, not ZnSe) and it played a role in Nichia's dramatic rise to Big 5 status in the LED industry. The patent is especially interesting because of how far it traveled (not that this is necessarily uncommon for patents nowadays). The first diagram below illustrates the connections between the '839 patent as found within the USPTO database. The patent was brought over from an original set of Japanese patent applications and therefore claims priority back to those JP applications (blue rectangles) for the various portions of its claims. Through a series of divisionals (based on the decision by the USPTO examiner that the claims in the US application actually refer to more than one "invention"), the original application resulted in a total of 8 US patents with the last one issuing in 2004.
 Zooming out from the USPTO perspective, however, we can see that this was not the extent of the patent's travels. By looking into the patent via the JPO, KPO, and the EPO's patent family search tool, we see the relationship between the original Japanese applications (along with the 5 patents they resulted in) as well as the 2 EU (taken into 2 German) and 4 Korean patents in its family.
 Finally, instead of taking the set of patents in a structural-style map, we can also see them over time in a way that gives insight into when Nichia's patent practitioners were working on which parts of this set of applications. The process viewed in this way reminds us how difficult it is for a patent that travels this far to be dealt with by a single patent practitioner. From the time of the first application (let alone the original technological work) in 1992 to the issuance of the final patent in its family in Korea in 2006 much had changed in Nichia. Nakamura, for instance, had long left the company for UC Santa Barbara, been sued by the company for trade secret theft, and won his own lawsuit over Nichia's failure to pay him adequate compensation for the value his technological work had brought to them.
Part of the take away from all of this is that we need to understand "property" not just in terms of what is owned or claimed, but also in terms of time. Patents especially are a sort of property with a very strange sort of relationship with time. Moreover, this is more than the difficulty of figuring out for certain when a patent's term is up. When I would ask my patent engineer informants how you could tell if a patent was a good one or not, they jokingly told me that you don't know until you use it in court, and win. Then how do you know if it has value? Well, you don't know that until it is used and upheld again. While the issuance of a patent (in most cases) means that the patent has already under gone (and passed) an examination to prove that it is "new," the interpretation of the patent law (in the US especially) may change several times over the course of the patent's life as the Supreme Court (or the Court of Appeals for the Federal Circuit) attempts to hone its guidance on what can and cannot be patented (and on a whole range of related rules and procedures). The value of a patent is thus predicated on a conditional future where it gets used and gets upheld. Yet, in the courtroom, the testimony that is most important for its validity (and thus its having or not having any value at all) is focused on a reconstructed past point in time where the technology was "invented" and the court's evaluation of the "state of knowledge of the art" at that time. Within itself, then, the patent references both past and future even as the process of claiming it may still be ongoing in jurisdictions in other parts of the world. Property is a process, not a state and understanding it takes time.
 


Creative Commons License
These diagrams by Matthew West are licensed under a
 
 

Friday, March 22, 2013

Mapping Taiwan's LED Industry

A diagram of past lawsuits and public cross-licensing activities within the global LED industry as it relates to Taiwanese LED companies (circa 2011). Taiwanese firms are represented in pink. Selected ownership relationships are represented by overlapping edges or more roughly by relative positions. Sources included numerous news releases and Satoshi Ookubo's previous compilation at Nikkei Electronics Asia, http://techon.nikkeibp.co.jp/article/HONSHI/20061130/124754/.


Creative Commons License
This diagram by Matthew West is licensed under a
 
 

TIN Anthropology

anthropological musings on property, piracy, materiality, and knowledge